researchers-to-deploy-ai-to-identify-lost-names-of-holocaust-victims-may-add-5-million-names-by-2029

Researchers to deploy AI to identify lost names of Holocaust victims, may add 5 million names by 2029

 • 69459 переглядiв

The World Holocaust Remembrance Center (Yad Vashem) is working to identify the names of victims who were previously lost among 230 million text documents, over the next few years Yad Vashem hopes to cover more than 5 million names, reports UNN with reference to Times of Israel.

Details

The World Holocaust Remembrance Center is working to identify the names of victims who were previously lost among 230 million text documents - in the coming years, the museum, whose name means "memorial and name" in Hebrew, will make greater use of artificial intelligence to identify previously unidentified victims.

It is noted that AI allows sorting and triangulating millions of documents to identify forgotten information. The software can make connections with Yad Vashem's vast archive that a human employee would not have the time or ability to make.

Yad Vashem spokesperson Simmy Allen said that the museum, whose name means "memorial and name" in Hebrew, will make greater use of artificial intelligence.

AI helps us by reviewing the evidence and removing names that have gone unnoticed until now. Until recently, we relied heavily on human intervention, but this process is extremely time-consuming

- Allen said.

"Over the next three to five years, Yad Vashem hopes to cover more than 5 million names. However, the further we move away from the events of the Holocaust, our work becomes much more difficult," said Alexander Avram, director of the Shoah Yad Vashem name recovery project.

Before the AI model could be deployed at Yad Vashem, the researchers had to determine how the names on their wartime lists could be sorted by the AI, Roth said.

Our initial step was to identify and understand the NER (Named Entity Recognition) and Relation Extraction database. To further improve the reliability and accuracy of [our model], we integrated significant logical collaboration by working closely with content experts from Yad Vashem

- says Shira Roth from the Information Technology Department.

AddendumAddendum

An award-winning artificial intelligence tool developed by Roth and her colleagues has produced remarkable results in the case of a survivor previously known only as Svailakh, a veterinarian from Lithuania.

"Svailakh was married to Eta Grudzanski, who is believed to have been killed in 1941 in Yurbarkas," Avram said. The main link to the Svailach name is the testimony of survivor Menachem Lichtenstein, who spoke of a man named Peter Svailach who was married to a woman named Ethel. These testimonies also relate to six other members of the Grudzinski family, most likely Ethel's relatives," noted Shoah Yad Vashem, director of the name recovery project.

According to him, Svailakh and his wife were killed by “locals” in the country. No one ever filled out the “Testimony Pages” for the couple's relatives, but the artificial intelligence tool Yad Vashem linked the Grudzinski family to Svailakh through Liechtenstein's testimony.

Thanks to [AI], these names can now be added to the database so that we can recover individuals from the oblivion of the Holocaust, as well as give a better idea of who they were before they became victims of anti-Semitism

- Abram said.

Recall

On the occasion of International Holocaust Remembrance Day, the exhibition project Live at Babyn Yar National Historical and Cultural Reserve opened. This was reported on the website of the Ministry of Culture and Information Policy of Ukraine.

Ihor Telezhnikov

News of the World

    Popular

    Ukrainian “Shchedryk” shines in Chanel's holiday campaign

     • 31344 переглядiв

    Billie Eilish is named Apple Music Artist of the Year

     • 73940 переглядiв

    S.T.A.L.K.E.R. 2 reaches one million sales in less than half a week after release

     • 78599 переглядiв

    Transparency or a screen: how ARMA transfers assets into management

     • 110901 переглядiв

    News by theme